The association between mineralised tissue formation and the mechanical local in vivo environment: Time-lapsed quantification of a mouse defect healing model

Abstract

An improved understanding of how local mechanical stimuli guide the fracture healing process has the potential to enhance clinical treatment of bone injury. Recent preclinical studies of bone defect in animal models have used cross-sectional data to examine this phenomenon indirectly. in this study, a direct time-lapsed imaging approach was used to investigate the local mechanical strains that precede the formation of mineralised tissue at the tissue scale. The goal was to test two hypotheses: 1) the local mechanical signal that precedes the onset of tissue mineralisation is higher in areas which mineralise, and 2) this local mechanical signal is independent of the magnitude of global mechanical loading of the tissue in the defect. Two groups of mice with femoral defects of length 0.85 mm (n = 10) and 1.45 mm (n = 9) were studied, allowing for distinct distributions of tissue scale strains in the defects. The regeneration and (re)modelling of mineralised tissue was observed weekly using in vivo microcomputed tomography (micro-CT), which served as a ground truth for resolving areas of mineralised tissue formation. The mechanical environment was determined using micro-finite element analysis (micro-FE) on baseline images. The formation of mineralised tissue showed strong association with areas of higher mechanical strain (area-under-the-curve: 0.91 ± 0.04, true positive rate: 0.85 ± 0.05) while surface based strains could correctly classify 43% of remodelling events. These findings support our hypotheses by showing a direct association between the local mechanical strains and the formation of mineralised tissue.

Reducing variance in a mouse defect healing model: Real-time Finite Element Analysis allows homogenization of tissue scale strains

“…In figure 4b&c, we can see the difference from a delicate trusslike structure to a more solidified cortical structure within the same bone via large changes in median strain. While the majority of mice for this model heal quickly and well, other models can include larger defects that require more time to bridge 16 , leaving the start date of loading in doubt. Even in the presented data, where the fracture gaps are relatively small, several mice displayed low bone volume at week 3 and hence had less structure to absorb load.…”

“…Here we analysed mice with the same fixation systems, same sex, operated on by the same surgeon in the same environment, ideally leading to a best-case scenario for external factors that could drive healing differences, yet variances remained. Even though it is well established that tissue level strains are the main driver behind remodelling and healing 2,16 the vast majority of mechanical interventions do not aim to target particular strains within the tissue directly.…”

“…For example, even within studies, defect size variance can be large, often displaying a standard deviation of over 10% of the nominal size in osteotomy based experiments [12][13][14] . In addition, even though it is known that the resulting strain from the initial size of the fracture gap 12,[15][16][17][18] is influential in the outcome of the healing process, this geometric information is regularly left either unreported, with no size or geometric description of the defect presented [19][20][21] , or underreported with only the nominal size being presented [22][23][24] .…”

“…Additionally, effects attributed to interventions such as pharmacological agents could be obscured due to inconsistent mechanical environments within and between groups. As loading models are often used to study the effects of mechanical stimulation on physiological processes in bone such as remodelling and healing, controlling the local and global stimuli acting on the tissue would improve the validity of such investigations 16 .…”

“…In this study, we analyse the inter-individual and temporal variance of the mechanical environment using image-based finite element analysis 16 during the bridging and post-bridging phase under a 10 N load and, due to individual differences seen within groups, propose and apply a novel methodology of adapting loading conditions to the individual bone geometry within an in vivo mouse femur osteotomy model. We term this method real time Finite Element (rtFE) adaptive loading.…”